[Domell]

Originally Posted by no-X

We should get some info much sooner. There are some NDA slides flying around, which will be public very soon

Things are not looking good according to Kyle@Hardocp who has said that we won`t see GT300 in retail in following 4-6 months.

[no-X]

Maybe it's the cause of spreading these GPGPU slides... I have no idea what's nVidias' current strategy.

[Arty]

Originally Posted by Ailuros

The rumor in that former link was speaking of a 512SP/384bit chip with the 512bit chip following later on. If true that would suggest a high end chip with castrated ROP partitions and not a performance GPU.

Albeit frankly yes ROPs might take up quite some die area but I can't figure out why theoretically 6 out of 8 partitions would make such a big difference to yields.

A 512b sku down to 384b makes sense only from a (lower cost) PCB perspective I guess. But I still dont understand how that helps Nvidia make it to the market quicker than the 512b sku? The highest priority right now is time to market, performance I'm guessing should be there unless something went horribly wrong.

Originally Posted by Ailuros

Of course would they have to have a significant lead; I just don't know or pretend to know (which is quite a popular trend lately during silly season *cough*) what the case exactly is. Now let's give the former rumor and the above thought a nice logical chain to follow: it sounds like a relatively bandwidth starved solution to me which could be good enough to surposs the competition's performance GPU by a good enough but not the best enough margin.

Good enough doesnt really sound good to be honest. R520 came three months later with a 10-15% boost, which quite simply was not enough.

Originally Posted by Ailuros

The maximum peak double precision throughput of a 5870 is at 544GFLOPs/s according to AMD itself.

Which would theoritically quite a leap from Cypress, granted but I recall AMD having a similar advantage current generation and yet it didnt pay off. I dont think this is anywhere as critical to boost the sales if the gaming perf isnt there.

Originally Posted by Ailuros

The scenario isn't even plausible from the get go in my mind. It could be some el weirdo test/dev sample like you can find from about any IHV before each launch, it could be a chip that is meant for a dual chip SKU, it could be some false translation, or Lord knows what else. Shall I remind you how specs for Cypress fluctuated until they finally came to the light of day? It nearly sounded like a rubber chip that you can stretch into all 4 directions.

So the die area D12U is supposed to consume varies between 452 and 576mm2.

How about the average of the two? 500mm2.

Originally Posted by Ailuros

Like the one from AMD for the Evergreen launch or an even older one:

http://users.otenet.gr/~ailuros/evergr.pdf

Wasnt this faked by some user who later owned up to it?

Originally Posted by Domell

Things are not looking good according to Kyle@Hardocp who has said that we won`t see GT300 in retail in following 4-6 months.

Link please? Next up Kyle would have got this information from a faulty translator.

[digitalwanderer]

Originally Posted by SiliconAbyss

They're fluffy AND juicy.

Ooooh, I like those kind. Feel free to PM them to me.

Originally Posted by Domell

Things are not looking good according to Kyle@Hardocp who has said that we won`t see GT300 in retail in following 4-6 months.

Got a link? Not that I doubt you, I just hate hunting for it.

Originally Posted by no-X

Seems they aimed much effort at nonexistent markets... Anyway, slides are one thing, reality another

CUDA isn't an non-existent market, it's just a hella lot smaller than the gaming market.

[Sinistar]

http://www.hardforum.com/showpost.ph...&postcount=362

http://www.hardforum.com/showpost.ph...&postcount=388

[Domell]

Here you are

http://www.hardforum.com/showpost.ph...&postcount=362

[digitalwanderer]

OMG, sorry for the topic combo breaker...but at one of the links from one of these posts I found this in the thread (and the original poster linked/quoted Arty which I really loved):

Nvidia has been VERY tight lipped about GT300 and for good reason I can assure you, these specs are within what I have been hinted towards from my Nvidia rep.

She is very excited about this product, the way she has been talking is that Nvidia is going to let ATi get comfortable with all the good press they are getting, and then they are going to drop the hammer with an amazing product line that is going to not only steal ATI's thunder, but put them to shame as well.

The thing she said that got me the most excited was exactly this "Nvidia has held absoultely NOTHING back this go around, they are giving GT300 Enthusiast cards everything they've got"

Anyone know any female nVidia reps? I don't off hand.

Thanks for the [H] links guys, gonna go check them now and try and figure out if nVidia is holding some killer cards in their hand or bluffing themselves into a hole.

[ChrisRay]

Originally Posted by digitalwanderer

OMG, sorry for the topic combo breaker...but at one of the links from one of these posts I found this in the thread (and the original poster linked/quoted Arty which I really loved):


Anyone know any female nVidia reps? I don't off hand.

Thanks for the [H] links guys, gonna go check them now and try and figure out if nVidia is holding some killer cards in their hand or bluffing themselves into a hole.

I know several

[trinibwoy]

Originally Posted by fellix

GPU TechCon schedule

Webcasts! Nice, didn't expect that.

Originally Posted by no-X

We should get some info much sooner. There are some NDA slides flying around, which will be public very soon

NDA slides for a card launching in 6 months? That's a long time to keep a secret

[no-X]

GPGPU related. Nothing about 3D performance or architecture. Seems nVidia want to convince us, that the product is more MORE MORE than only a gaming accelerator. I expect it shall create WOW effect and (together with the number of SPs) convince customers to wait.

Are you ready?

[psurge]

Too bad about the lack of arch details. Still, I'm thinking b3d should probably raise the limit on the number of people that a PM can be sent to simultaneously.

anytime no-X

[nAo]

Ehi, my pm folder is empty!

[MfA]

Originally Posted by no-X

MORE MORE than only a gaming accelerator.

My entire computer is very little more than a gaming accelerator ... hell if we ignore FaH&Co for a moment gaming is probably the only thing which stresses the computer for 99% of us.

[SiliconAbyss]

Originally Posted by digitalwanderer

Ooooh, I like those kind. Feel free to PM them to me.

I would love to, but sorry don't have access to said slides. I know I just broke your heart.

The thing she said that got me the most excited was exactly this "Nvidia has held absoultely NOTHING back this go around, they are giving GT300 Enthusiast cards everything they've got"

You mean Nvidia has been holding back until now?

[Scali]

Originally Posted by SiliconAbyss

You mean Nvidia has been holding back until now?

They've been holding back on DX10.1, 40 nm and GDDR5 at least

[Ailuros]

Originally Posted by Arty

A 512b sku down to 384b makes sense only from a (lower cost) PCB perspective I guess. But I still dont understand how that helps Nvidia make it to the market quicker than the 512b sku? The highest priority right now is time to market, performance I'm guessing should be there unless something went horribly wrong.

As I said Lord knows where that one came from. I'd be even willing to believe that someone couldn't find any other reasonable explanation for the infamous GTX280 1.5GB that appeared in some weird roadmaps.

Good enough doesnt really sound good to be honest. R520 came three months later with a 10-15% boost, which quite simply was not enough.

Since none of us knows where it's performance ballpark lies it doesn't help coming up with any percentages either.

Which would theoritically quite a leap from Cypress, granted but I recall AMD having a similar advantage current generation and yet it didnt pay off. I dont think this is anywhere as critical to boost the sales if the gaming perf isnt there.

Why didn't it pay off for AMD exactly? I don't follow market shares that often but I recall at some point reading that AMD's share there grew in one quarter from 8% to 12%. And before the dreaded "so what" remark comes along that's a 50% increase in market share in a low volume ultra high margin market.

How about the average of the two? 500mm2.

That's the sad reality that the truth usually lies in the middle. Take neliz's prediction that it'll be by only 20% than X as worst case scenario and NV's own bullshit exaggerations stir them slightly and there you have your coffee coloured performance estimate

[SiliconAbyss]

Originally Posted by Scali

They've been holding back on DX10.1, 40 nm and GDDR5 at least

Lol touché

[neliz]

Originally Posted by Ailuros

That's the sad reality that the truth usually lies in the middle. Take neliz's prediction that it'll be by only 20% than X as worst case scenario and NV's own bullshit exaggerations stir them slightly and there you have your coffee coloured performance estimate

Are you hinting that GTX 380 boards will be brown? HURRAH!

Now if it could only smell like coffee

we just have to wait a few more weeks to get the first pics and numbers out.

[Ailuros]

What GTX 380? :P

[DegustatoR]

Originally Posted by Ailuros

What GTX 380? :P

You mean we should start calling it GF380? 8)

[Unknown Soldier]

Originally Posted by neliz

Are you hinting that GTX 380 boards will be brown? HURRAH!

Now if it could only smell like coffee

we just have to wait a few more weeks to get the first pics and numbers out.

It'll most probably cook your coffee for you

Originally Posted by DegustatoR

You mean we should start calling it GF380? 8)

I thought it was G380

[Ailuros]

G-Force 380 ya know the squirl with the fatter butt *snicker*

[DegustatoR]

Originally Posted by Unknown Soldier

I thought it was G380

No no that's the low end OEM mobile part with DX10.1 support only.

[Jawed]

Originally Posted by DegustatoR

You mean we should start calling it GF380? 8)

Yeah, Fermi not Tesla, eh?

Jawed

[Vincent]

Fermi---> ( Server Grade GPU for HTPC )